Deformable convolutions add 2D offsets to the regular grid sampling locations in the standard convolution. It enables free form deformation of the sampling grid. The offsets are learned from the preceding feature maps, via additional convolutional layers. Thus, the deformation is conditioned on the input features in a local, dense, and adaptive manner.
Source: Deformable Convolutional NetworksPaper | Code | Results | Date | Stars |
---|
Task | Papers | Share |
---|---|---|
Semantic Segmentation | 24 | 10.00% |
Object Detection | 24 | 10.00% |
Super-Resolution | 13 | 5.42% |
Optical Flow Estimation | 10 | 4.17% |
Instance Segmentation | 9 | 3.75% |
Video Super-Resolution | 8 | 3.33% |
Image Segmentation | 8 | 3.33% |
Image Classification | 6 | 2.50% |
Medical Image Segmentation | 6 | 2.50% |
Component | Type |
|
---|---|---|
🤖 No Components Found | You can add them if they exist; e.g. Mask R-CNN uses RoIAlign |